A Control Algorithm to Improve Target Tracking in Individuals with and without Cerebral Palsy

Harshal Mahajan1-2 ; Brad E. Dicianno, MD1-3 ; Geoffrey J. Gordon, PhD4 ; Rory A. Cooper, PhD1-3 ; Hongwu Wang1-2 ; and Sara Sibenaller1-2
1 Department of Rehabilitation Science and Technology, University of Pittsburgh
2 Human Engineering Research Laboratories (HERL), Highland Drive VA Medical Center, Pittsburgh PA
3 Department of Physical Medicine and Rehabilitation, University of Pittsburgh Medical Center
4 Dept. of Machine Learning, Carnegie Mellon University, Pittsburgh, PA


People with Cerebral Palsy have difficulty operating conventional control interfaces due to abnormal posture and coordination, increased tone, and involuntary movements. This may impair such necessary functional activities as power mobility, computer access, or communication via augmentative devices and may have a significant impact on quality of life. The purpose of this paper is to present preliminary data on an advanced algorithm applied offline to previously collected data from individuals with and without CP who performed computer target tracking tasks using a force sensing control.  Our aim was to test the utility of using the algorithm offline to correct target tracking trajectories when users “overshoot” or miss a target.  The algorithm significantly improved trial times for all subjects, especially for the subject with CP.  On average, his trials were shortened from over a minute to just a few seconds.  These results have exciting potential application for improving on-screen computer tasks, webpage navigation, and augmentative communication.  We expect even more robust results when the algorithms are used online during subject trials.  Future work will focus on machine learning techniques that aim to predict subjects’ desired trajectories in tasks such as power wheelchair driving in which subjects’ desired trajectories are more difficult to predict.


Cerebral Palsy, Computer-user Interface, Joysticks, Machine Learning, Wheelchairs


The funding for this research was provided by the Rehabilitation Medicine Scientist Training Program NIH K12 Award (K12HD01097).

Brad E. Dicianno, MD, Human Engineering Research Labs, VA Pittsburgh Healthcare System, Pittsburgh, PA 15206, Phone: (412) 365-4850, Email: diciannob@herlpitt.org